Merge pull request #14 from LokiMetaSmith/fix-cpu-ddp-init

Fix hardware detection for AMD ROCm and single-process CPU crashes

output.txt

@@ -1,34 +0,0 @@
-
-                                                       █████                █████
-                                                      ░░███                ░░███
-     ████████    ██████   ████████    ██████   ██████  ░███████    ██████  ███████
-    ░░███░░███  ░░░░░███ ░░███░░███  ███░░███ ███░░███ ░███░░███  ░░░░░███░░░███░
-     ░███ ░███   ███████  ░███ ░███ ░███ ░███░███ ░░░  ░███ ░███   ███████  ░███
-     ░███ ░███  ███░░███  ░███ ░███ ░███ ░███░███  ███ ░███ ░███  ███░░███  ░███ ███
-     ████ █████░░████████ ████ █████░░██████ ░░██████  ████ █████░░███████  ░░█████
-    ░░░░ ░░░░░  ░░░░░░░░ ░░░░ ░░░░░  ░░░░░░   ░░░░░░  ░░░░ ░░░░░  ░░░░░░░░   ░░░░░
-
-Overriding: depth = 2
-Overriding: device_batch_size = 2
-Overriding: max_seq_len = 128
-Overriding: num_iterations = 10
-Overriding: run = dummy
-Autodetected device type: cpu
-2025-11-22 17:01:10,440 - nanochat.common - INFO - Distributed world size: 1
-/app/nanochat/tokenizer.py:397: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
-  token_bytes = torch.load(f, map_location=device)
-Vocab size: 265
-num_layers: 2
-model_dim: 128
-num_heads: 1
-num_kv_heads: 1
-Tokens / micro-batch / rank: 2 x 128 = 256
-Tokens / micro-batch: 256
-Total batch size 524,288 => gradient accumulation steps: 2048
-Number of parameters: 461,056
-Estimated FLOPs per token: 2.956032e+06
-Using user-provided number of iterations: 10
-Total number of training tokens: 5,242,880
-Tokens : Params ratio: 11.37
-Total training FLOPs estimate: 1.549812e+13
-Scaling the LR for the AdamW parameters ∝1/√(128/768) = 2.449490

pyproject.toml

@@ -47,7 +47,8 @@ python_functions = ["test_*"]
 [tool.uv.sources]
 torch = [  
     { index = "pytorch-cpu", extra = "cpu" },  
-    { index = "pytorch-cu128", extra = "gpu" },  
+    { index = "pytorch-cu128", extra = "gpu" },
+    { index = "pytorch-rocm", extra = "amd" },
 ]
 
 [[tool.uv.index]]  
@@ -60,6 +61,11 @@ name = "pytorch-cu128"
 url = "https://download.pytorch.org/whl/cu128"  
 explicit = true
 
+[[tool.uv.index]]
+name = "pytorch-rocm"
+url = "https://repo.amd.com/rocm/whl/"
+explicit = true
+
 [project.optional-dependencies]  
 cpu = [  
     "torch>=2.8.0",  
@@ -67,11 +73,15 @@ cpu = [
 gpu = [  
     "torch>=2.8.0",  
 ]  
+amd = [
+    "torch>=2.8.0",
+]
   
 [tool.uv]  
 conflicts = [  
     [  
         { extra = "cpu" },  
-        { extra = "gpu" },  
+        { extra = "gpu" },
+        { extra = "amd" },
     ],  
 ]  

speedrun.sh

@@ -23,8 +23,21 @@ mkdir -p $NANOCHAT_BASE_DIR
 command -v uv &> /dev/null || curl -LsSf https://astral.sh/uv/install.sh | sh
 # create a .venv local virtual environment (if it doesn't exist)
 [ -d ".venv" ] || uv venv
+
 # install the repo dependencies
-uv sync --extra gpu
+# Detect hardware to install the correct torch version
+if command -v nvidia-smi &> /dev/null; then
+    echo "NVIDIA GPU detected. Installing CUDA dependencies..."
+    EXTRAS="gpu"
+elif [ -e /dev/kfd ]; then
+    echo "AMD GPU detected. Installing ROCm dependencies..."
+    EXTRAS="amd"
+else
+    echo "No dedicated GPU detected. Installing CPU dependencies..."
+    EXTRAS="cpu"
+fi
+uv sync --extra $EXTRAS
+
 # activate venv so that `python` uses the project's venv instead of system python
 source .venv/bin/activate